Effects of simulated low-temperature thermal remediation on the microbial community of a tropical creosote contaminated soil

Abstract

In the search for more sustainable remediation strategies for PAH-contaminated soils, an integrated application of thermal remediation and bioremediation (TEB) may allow the use of less impacting temperatures by associating heating to biological degradation. However, the influence of heating on soil microbiota remains poorly understood, especially in soils from tropical regions. This work investigated the effects of low-temperature heating on creosote-contaminated soil bacteria. We used culture-dependent and 16 S rRNA sequencing methods to compare the microbial community of soil samples heated to 60 and 100 ^oC for 1 h in microcosms. Heating to 60 °C reduced the density of cultivable heterotrophic bacteria compared to control soil (p < 0.05), and exposure to 100 °C inactivated the viable heterotrophic community. Burkholderia-Caballeronia-Paraburkholderia (BCP) group and Sphingobium were the predominant genera. Temperature and incubation time affected the Bray-Curtis dissimilarity index (p < 0.05). At 60 °C and 30 days incubation, the relative abundance of Sphingobium decreased and BCP increased dominance. The network of heated soil after 30 days of incubation showed fewer nodes and edges but maintained its density and complexity. Both main genera are associated with PAH degradation, suggesting functional redundancy and a likely potential of soil microbiota to maintain biodegradation ability after exposure to higher temperatures. We concluded that TEB can be considered as a potential strategy to bioremediate creosote-contaminated soils, allowing biodegradation in temperature ranges where thermal remediation does not completely remove contaminants. However, we recommend further research to determine degradation rates with this technology.

Introduction

Contaminated sites pose a common issue in our society, representing a significant threat within the framework of the One Health model. Pollutants, such as polycyclic aromatic hydrocarbons (PAHs), have implications for the health of environmental compartments such as the atmosphere, water, and soil. Their presence contributes to biodiversity loss, as well as the degradation of ecosystem functions and services [1, 2]. To remediate and manage these pollutants, a variety of technologies are available, which rely on chemical, physical, or biological principles, often involving interactions between these methods [3]. Recently, there has been growing interest in new approaches to remediation, such as Nature-based Solutions (NbS), particularly to achieve more sustainable methods [4]. NbS aims to minimize secondary impacts, maximize benefits, and provide resilient solutions [4].

Bioremediation, harnessing the metabolic capabilities of microorganisms or plants to degrade pollutants, aligns with the principles of NbS [5, 6]. However, its effectiveness depends on severall factors including contaminant bioavailability, toxicity, and the interplay of physical-chemical and biological aspects (e.g., nutrients, oxygen, temperature, pH, humidity, microbial abundance, and ecological interactions) [7, 8]. Creosote is a pollutant composed of high molecular weight (HMW) PAHs and is known for its toxicity, recalcitrance, and limited bioavailability in soil, exemplifies a challenging pollutant for bioremediation efforts [9]. Achieving remediation goals for such recalcitrant compounds via bioremediation can be a lengthy process, influenced by site-specific characteristics [10–13].

Therefore, alternative technologies can be employed to address the limitations in bioremediation efficacy. Among the remediation technologies available, thermal desorption (TD) emerges as a viable option, particularly for its high efficiency over a short timeframe [14–17]. TD can be categorized into low temperature thermal desorption (LTTD) and high temperature thermal desorption (HTTD), with LTTD operating within the range of 100 °C to 300 °C, and HTTD reaching temperatures from 300 °C to 500 °C [17]. However, the energy-intensive nature of TD, particularly at higher temperatures, raises concerns about its sustainability, as it can lead to atmospheric emissions and soil property damage [18–21]. Heating the soil during TD treatments, especially at 100 °C, can alter soil organic matter (SOM), while temperatures of 200 °C to 250 °C can induce changes in soil texture, mineralogy, and pH, varying across soil types [21, 22]. Heating also impacts soil microbiota, with studies indicating a decrease in the abundance of bacteria and fungi with increasing temperatures up to 300 °C, particularly evident from 200 °C onwards [23]. Conversely, LTTD treatments tend to promote soil recolonization and establishment of a more diverse and heat-tolerant microbial community [22–25]. However, the majority of studies have been conducted with soils from temperate regions, highlighting a gap in research on soils from tropical areas.

The combined use of thermal and bioremediation, called thermally enhanced bioremediation (TEB) is increasingly being considered as an alternative to overcome the disadvantages of both technologies [26, 27]. The bioremediation is not so effective for HMW PAHs, on the other hand the HTTD has problems with energy use and sustainability. Usually, LTTD at 80 to 100 °C is used to treat soils contaminated with LMW PAHs [21], which may be favourable for the implementation of TEB, as microbiota will be less affected by temperature at this range. In addition, LTTD uses less energy and has less impact on soil properties, and, although it shows lower efficiency in removing less volatile compounds such as the HMW present in creosote, it may stimulate bioremediation with native microbiota, increasing the bioavailability of pollutants [27, 28]. However, our understanding on how heating affects not only the survival but also diversity in soil microbiomes still needs to improve to allow a more effective application of this combined remediation strategy [27], especially in soils from tropical areas. Heated contaminated soils present changes in the taxonomy, recovery rate and functional redundancy depending on the temperature, contaminants, and time after heating [22, 29, 30].

In the pursuit of more sustainable and novel remediation strategies that integrate cost-effectiveness and soil health, the utilization of combined methods can optimize benefits while minimizing impacts [15, 31, 32]. In this context, a comprehensive understanding of microbiome behavior interacting with other remediation methods is crucial. In this study, we simulated LTTD (60 and 100 °C) heating of soil from a creosote-contaminated tropical area to assess its impact on microbiota diversity and their ability to survive and recolonize the soil, using both culture-dependent methods and 16 S rRNA sequencing.

Materials and methods

Soil samples

Soil samples were collected in a contaminated site in the west of the city of São Paulo, São Paulo State, Brazil, previously used for cargo railway operations and wood treatment with creosote. Contamination has persisted for about 30 years in the area [33]. A 2019 report identified naphthalene (35.6567 mg.Kg^− 1), dibenzofuran (23.196 mg.Kg^− 1), and phenanthrene (7.322 mg.Kg^− 1) as the main PAHs in the soil [34]. Approximately 5 kg of soil was collected using a hand auger at a depth of 40 to 60 cm. Subsamples were taken, transferred to sterile bags, homogenized, and sieved through a 2 mm sieve. Thirty-six microcosms were created by adding 30 g of soil to 50 mL conical tubes. All manipulations were conducted in a laminar flow hood using autoclaved or 70% ethanol-cleaned materials to prevent external contamination of the samples.

Thermal treatment of soil microcosms

Batches of 12 screw-capped microcosms were dry heated at 60 °C and 100 °C, for one hour in the dark and then incubated at 30 °C in a BOD incubator, also in the dark. Microcosms were weighed before and after heating, with water losses under 2% not requiring refilling. Triplicates were sampled immediately after heating (time zero) and after 10, 30, and 55 days of incubation. A control batch of unheated microcosms was also incubated at 30 °C in the dark. Samples were labeled by heating temperature (C for control, 60, or 100), incubation time (t0, t10, t30, and t55), and replicate number (r1, r2, or r3). At each incubation time, 1 g subsamples were analyzed using heterotrophic plate count analysis, and the remaining soil was stored at -20 °C for 16 S rRNA sequencing. The experimental design was performed based on a previous study [22], where the 60 and 100 °C were temperatures that affected the bacterial community, but showed some ability to recover [22].

Physicochemical analysis

Soil samples (30 g each) heated at 60 °C and 100 °C, along with an unheated control group kept at room temperature, were analyzed for physicochemical properties and micronutrients. Boron (B) was extracted with boiling water and measured via colorimetry. Copper (Cu), Zinc (Zn), Manganese (Mn), and Iron (Fe) were extracted with DTPA and analyzed using atomic absorption spectrophotometry. Sodium (Na) was extracted with Mehlich 1 and quantified by flame photometry. Silicon (Si) was extracted with 0.01 mol/L calcium chloride solution and determined colorimetrically. Soil pH was measured in a 0.01 mol/L CaCl₂ solution. Phosphorus (P) and Potassium (K) were extracted with ion exchange resins and quantified by colorimetry and atomic emission spectrophotometry, respectively. Calcium (Ca) and Magnesium (Mg) were extracted with potassium chloride and determined using atomic absorption spectrophotometry. The SMP buffer was used to extract H + Al, and Cation Exchange Capacity (CEC) was calculated by summing all exchangeable cations. Exchangeable bases (SB) were determined by summing all exchangeable cations except H⁺ and Al³⁺, while the base saturation percentage (BS) was calculated from the exchangeable bases. Granulometric analysis was conducted on the unheated soil according to Raij et al. 2001 [35].

Heterotrophic plate count

Densities of viable culturable heterotrophic bacteria were used as an indicator of the ability of the soil microbial community to resist the increased temperature impact and recolonize the soil after heating. Counting was performed by the small drop assay [36] in subsamples of 1 g of soil previously vortexed for 2 min in 10 mL sodium pyrophosphate aqueous solution (0.1%) and settled for 30 min. Plates were incubated at 30 °C for 36 h before counting.

DNA extraction and sequencing

The genomic DNA of 0.5 g of soil from microcosms was extracted with the DNeasy PowerSoil Pro Kit (QIAGEN) following manufacturer’s instructions. Before extraction, 500 µL of 80 mg.g^− 1 sterile (autoclaved) powdered milk (ACUMEDIA) prepared as described in de Carvalho et al. (2016) was added to the sample. The milk competes for adsorption sites in soil, enhancing DNA extraction efficiencies. PCR amplification of the V4 hypervariable region of 16 S rRNA gene was carried out with the primers 515 F and 806R and conditions described in Caporaso et al. (2011) [37] with 40 cycles. The sequencing libraries were generated using Ultra DNA library Pre ^® Kit for Illumina, with quality assessed on the Qubit 2.0 Fluorometer (Thermo Scientific). The sequencing was performed on the Illumina HiSeq2500 platform to generate 250 bp paired-end reads by GenOne Biotech.

Bioinformatic analysis

Initial removal of the primers was performed with Cutadapt in Qiime2 (version 2020.6 [38]). Quality of sequences was verified with FastQC (0.11.9 [39]). All samples were within the parameter limit (phred score > 30), so no filter was applied to the samples. Chimeras were removed and the denoising was performed with DADA2 (1.14 [40]). Silva’s database (Release 138 [41]), was used for taxonomic assignment and BLAST (version 2.11.0) used to confirm the assignment’s quality for the ASVs (Amplicon Sequence Variant). Mitochondrial and chloroplast sequences were removed with the Qiime2 (version 2020.6).

Statistical analysis

Diversity and statistical analysis were performed, and graphics were drawn by the Rstudio environment (4.3.2) with the main package microeco [42] and other needed packages contained in it. To test linear correlations the lm() function from the stats package was used; also, to compare total CFU between treatments a t-test was used. Besides that, a PERMANOVA with 1 × 10⁵ permutations was performed to evaluate the effects of the treatment in beta diversity metrics. In order to have a deeper understanding about how the temperature affects not only the composition of the bacterial communities, but also the co-ocurrence patterns we analysed the co-ocurrence networks. It was considered the networks in the early stages of the community after heating (0 to 10 days) and several days after the heating (30 to 55 days). Only samples heated at 60 °C were considered for this analysis because of the statistical need to have at least three samples to build a network. The networks were built using the Sparse Correlations for Compositional Data (SparCC) technique, which uses linear Pearson correlations and is a robust model [43–45]. Besides that, the Gephi software (v. 0.10.1; [46]) was used to calculate topological parameters and also visualize the networks. Network complexity and parameters were analysed considering an initial incubation period (0 to 10 days after heating) and the remaining period (10 days to 55 days after heating). The topological parameters from the network analysed were: number of nodes and edges; average degree; average path length; network diameter; cluster coefficient; density.

Results

Physicochemical analysis

The soil from the area is predominantly composed of clay, has low values of natural organic matter and is slightly acidic (Table 1). The micronutrients iron (Fe) and silicon (Si) decreased by up to 20 and 8% respectively with heating. In addition, copper and potassium showed an increase. The other physicochemical properties from the soil did not change drastically with the heating at 60 °C and 100 °C.

Table 1.

Physicochemical properties of the control and heated soil samples

Parameter	Unit	Control	Heated at 60 °C	Heated at 100 °C
Clay	%	55.2	-	-
Silt	%	3.2	-	-
Total sand	%	41.6	-	-
Organic Matter	g/dm³	7.1	6.8	7.1
pH	-	5.84	5.97	5.87
P	mg/dm³	< 7	< 7	< 7
Ca	mmolc/dm²	20.6	20.4	20.1
Mg	mmolc/dm³	3.2	3.1	3.0
K	mmolc/dm³	1.63	1.74	1.83
H + Al	mmolc/dm²	15.8	15.8	15.7
SB	mmol/dm³	25.3	25.2	24.9
CEC	mmol/dm³	41.1	41.0	40.6
BS	%	62	62	61
B	mg/dm³	1.04	0.93	0.98
Cu	mg/dm³	1.5	1.4	5.4
Fe	mg/dm³	6.2	5.4	5.0
Mn	mg/dm³	4	3.8	3.7
Na	mg/dm³	33	31	32
Si	mg/kg	9.7	9.0	8.9

Heterotrophic cultivable bacteria

In the control soil, populations of culturable heterotrophic bacteria increased throughout the incubation period, except at 30 days after incubation, ranging from 6,7 × 10⁶ to 1,2 × 10⁷ CFU.g^− 1 (Fig. 1). There were no significant differences between 60 °C and control when considering incubation time (Fig. 1A - linear model p > 0.05). However, there was a significant difference in total CFU, with microcosms heated to 60 °C having lower total CFU compared to the control (Fig. 1B - t-test p = 0.016). In addition, populations recovered over the incubation period, and after 30 days, cell densities were still lower but tended to approach those in the control microcosms.

Fig. 1.

(A) Linear regression of cellular densities (CFU.g^− 1) of heterotrophic bacteria over time for control and heated samples at 60 °C, and (B) Boxplot for cellular densities (CFU.g^− 1) of heterotrophic bacteria through 55 days of incubation for control and heated samples at 60 °C

Conversely, microcosms heated at 100 °C showed no heterotrophic cellular density at any incubation time, indicating that culturable bacteria were inactivated by heating under the experimental conditions.

Changes in the bacterial total diversity

Analysis of the 16 S rRNA gene sequences revealed a predominance of the Pseudomonadota phylum (Fig. 2). Bacillota were also abundant in microcosms Ct55r1, 60t30r2, and 100t30r1, while an increased abundance of Bacteroidetes was observed in microcosm 100t30r1.

Fig. 2.

Relative abundance of bacteria at phylum and genus level from control and heated samples. Similar colors indicate genera from the same phylum. The white portion of the bars correspond to the low abundance (< 0.1%) or unidentified families. Sample codes: C, control; 60, 100, heating temperatures; t0, t10, t30, t55, incubation times in days; r1, r2, r3, replicates

At the genus level, the most abundant Pseudomonadota were Burkholderia-Caballeronia-Paraburkholderia (BCP) and Sphingobium. The only Bacillota with higher abundance was Clostridium and for Bacteroidetes phylum was Muribaculaceae. The microcosm groups were divided into three main clusters based on the Bray-Curtis distance (Fig. 3). The first cluster (samples Ct55r2, 60t30r1, 60t55r1, 60t55r2, and 100t55r1) primarily contained microcosms incubated for 55 days, showing a predominance of the BCP group (over 50%) and the genera Escherichia-Shigella (between 10 and 20%). The second cluster grouped samples heated at 60 °C with up to 10 days of incubation and control samples incubated for up to 30 days, mainly represented by the genera Sphingobium (30%) and BCP (40%). The last cluster grouped samples Ct55r1, 60t30r2, 100t30r1, mainly composed of Clostridium_sensu_sricto_1 (12–49.5%) followed by BCP (18.4–22.7%), except for sample 100t30r1, which presented the genus Muribaculum (31.6%) as predominant.

Fig. 3.

Heatmap at genus level based on the microbial community similarity. The phylogenetic tree is represented on the left and cluster by similarity on top

Alpha diversity measurements (Fig. 4) indicated a decreasing trend with increasing incubation time in all treatments, suggesting a decrease in species abundance and evenness over time. Although there was no statistical difference between the linear regressions, the slopes for 60 °C treatments were higher, indicating a tendency for a more pronounced decrease in alpha diversity measures throughout the incubation period.

Fig. 4.

Linear regression for alpha diversity parameter: Chao1, Shannon and Simpson

Beta diversity indexes (Figs. 1S and 5) measured by Bray-Curtis, Jaccard, Unifrac, and Weighted Unifrac distances reinforced the grouping pattern of samples observed previously (Fig. 3). One group comprised samples from 0 to 30 days of incubation of control/60°C heated microcosms. The other two groups contained 30 and 55 days of incubation microcosms from all treatments: one containing three microcosms with a higher abundance of Clostridium and the other comprising the remaining microcosms from 30 to 55 days of incubation. A PERMANOVA analysis performed for Bray-Curtis and Jaccard distances revealed that temperature and time influenced diversity for Bray, whereas for Jaccard, only time had an impact (Table 2). For no beta diversity measure, the combined effect of temperature and time had an impact on the community.

Fig. 5.

Beta diversity distances in different samples. The distances used were Bray-Curtis, Jaccard, Unifrac and Weighted Unifrac. Samples are separated by temperature and incubation time

Table 2.

PERMANOVA Test with 1 × 10⁵ permutations results for Beta diversity parameters. Asterisk indicates statistically significant differences in p-values at standard confidence levels (0.95)

Variable	Parameter	p value	R²
Temperature	Bray-Curtis	0.045*	0.168
Incubation time	Bray-Curtis	0.007*	0.354
Temperature: Incubation time	Bray-Curtis	0.348	0.173
Temperature	Jaccard	0.538	0.097
Time	Jaccard	0.028*	0.280
Temperature: Incubation time	Jaccard	0.517	0.201

In order to understand the complexity of the microbial community before and after the heating, a co-occurrence network analysis was performed using SparCC (Fig. 6; Table 3). Comparing the control samples from the initial to the final time revealed a decrease in both nodes and edges, resulting in a lower node/edge ratio (0.171 to 0.193) and a reduction in the average degree (from 5.843 to 5.171). Additionally, there was an increase in the average degree path length (1.893 to 2.194), with the network diameter expanding from 5 to 7. Also, the clustering coefficient decreased (0.189 to 0.176), while the density increased (0.018 to 0.02). In contrast, the heated samples showed a more pronounced decrease in nodes and edges over time (node/edge ratio: 0.191 to 0.291). The average degree decreased from 5.241 to 3.435, followed by an increase in the average path length (1.995 to 2.523). The network diameter also increased by one, while both the clustering coefficient and density rose (from 0.165 to 0.176 and 0.015 to 0.02, respectively).

Fig. 6.

Network from control and heated samples from 0 to 10 days and 30 to 55 days. The network graphs were built with SparCC (SparCC value > 0.8 or < -0.8 and Pearson’s p < 0.01). Build in Gephi Software

Table 3.

Correlation and properties of the networks (p < 0.05)

Topological Parameters	Control 0–10 days	Control 10–55 days	60 °C 0–10 days	60 °C 10–55 days
Node/edge ratio	0.171	0.193	0.191	0.291
Nodes	325	258	349	177
Edges	1899	1334	1829	608
Average degree	5.843	5.171	5.241	3.435
Average path length	1.893	2.194	1.995	2.523
Network diameter	5	7	6	7
Cluster coefficient	0.189	0.176	0.164	0.176
Density	0.018	0.02	0.015	0.02

Discussion

In this study, heating at 60 °C and 100 °C slightly altered the soil’s properties without significantly affecting organic matter, but decreased iron content, likely due to oxidation [47]. Regarding the effects of temperature on microbial communities, Pingree and Kozibar (2019) [48] noted that thermal thresholds for organisms vary due to differences in conditions and soil types. Typically, bacterial thresholds range from 60 °C to 100 °C. In this study, 100 °C disrupted the bacterial community and consequently hindered PAH biodegradation processes that may be occurring in the soil. In contrast, bacterial communities exposed to 60 ^oC, despite suffering a decrease in culturable populations short after the heating, showed a trend to recover with time. However, heating resulted in changes in microbial diversity.

The genus Sphingobium was present in all samples, but in decreasing proportions with increasing incubation time and heating temperature. This genus is already known to degrade several PAHs [49, 50] including naphthalene. Therefore, its presence in the study site is expected and favours biodegradation. Although this genus is metabolically versatile and genetically adapted to different environmental conditions [50–52], the results indicated that it is sensitive to temperature, in contrast to BCP group, whose abundance increased along incubation time, becoming a predominant group in the microbiota of the microcosms.

The decreasing trend in Simpson index suggested that longer incubation times may lead to the dominance of specific groups within the communities. Besides that, PERMANOVA analysis revealed that both temperature and time significantly impacted Bray-Curtis dissimilarity, indicating differences in microbial community abundance. In contrast, temperature did not affect the Jaccard distance, likely because this index is presence/absence-based, showing that the taxonomic groups were similar in control and heated communities. The heatmap illustrated higher abundance of the BCP group in samples heated at 60 °C. Time also significantly influenced beta diversity, with BCP becoming dominant in samples over time.

Compositional changes from taxa to ASVs were evident in co-occurrence networks, reflecting the response of soil bacterial communities to environmental perturbations [53]. Both control and heated samples showed networks with lower average degree values, indicating reduced network co-occurrence relations. The average path length, which indicates the network’s ability to respond to environmental changes [54, 55], increased in soil heated at 60 °C after 30 days, possibly due to the loss of certain taxa. However, network density, which indicates complexity and correlates with interactions and energy requirements [56, 57], remained unchanged. According to the Domino hypothesis, less dense networks have lower maintenance costs and greater resilience [57]. Thus, despite heating disturbances, the network density remained stable after 30 days of incubation, suggesting that the system’s efficiency was not compromised due to its resilience and potential functional redundancy.

The observed changes in bacterial diversity and networks do not necessarily mean a negative impact on bioremediation potential of the soil in the studied area. BCP group has been reported to completely degrade naphthalene at 1 g/L [58] and to degrade an intermediate during phenanthrene degradation (2-hydroxy-1-naphthoic acid − 2H1NA) [59]. In addition, it has a large diversity of catabolic enzymes that can degrade pollutants, such as: toluene-2,3-dioxygenase and chlorocatechol-1,2-dioxygenase (degrade toluene); 1,2-dihydroxy-1,2-dihydronaphthalene dehydrogenase (NahB) (degrade naphthalene analogs) [60]; and several others specific and generic enzymes [58–62]. Besides that, the BCP group was reported to tolerate naphthalene in concentrations up to 10-fold higher than do other genera [58, 63–65] and to degrade pollutants such as triazophos and phenanthrene at higher rates than other genera [66, 67]). Burkholderia spp. is also reported to degrade several compounds and utilise PAHs as the sole carbon source [68].

In addition to the ability of degrading PAHs, BCP has high adaptability to a wide range of environments [58]. Their bigger genome (11.5 Mb and 2–4 chromosomes) confers metabolic advantages and BCP has also a higher number of insertion sequences that can help stimulate flexibility and genomic plasticity [69, 70]. Because of these genomic adaptations and advantages, several works registered BCP degrading pollutants under pH range from 3 to 9 [66–71]. This work also demonstrated that the genus BCP was able to survive under a heating temperature of 60 °C, unlike the genus Sphingobium, which showed a decrease in its abundance.

The compounds with higher concentration values present in the creosote (phenanthrene, acenaphthene, naphthene, fluoranthene) have boiling temperatures higher than 200 °C. Therefore, at temperatures of 60–100 °C the boiling point is not reached, but there will be an increase in the volatilization rate. The increase in temperature also increases vapour pressure [72]. Besides that, Henry’s constant can increase with the heating until it reaches a peak that will start to decrease with temperature [22]. These two properties influence the volatilization rate of a compound, and with the heating, the volatilization rate increases [73]. This can cause a slight decrease in concentration of the compounds present in the creosote, but not a full removal. The heating also increases the solubility and decreases the K_ow (octanol-water partition coefficient) [22, 74]. The solubility and K_ow are related to the sorption capability from a soil, which in this case causes the compounds to be less retained by sorption. Besides the sorption capability, the heating also decreases the viscosity from a compound [26, 27]. Therefore, it is not expected that heating temperatures of 60–100 °C will remove or degrade completely the PAHs in creosote, but these changes in compound characteristics, such as increase in the volatility, enhance the bioavailability from the contaminants.

In this context, the concurrent application of low-temperature thermal desorption and biodegradation is emerging as a less impactful alternative to the application of high-temperature thermal remediation, which is energy intensive, can disrupt soil properties, and may inactivate bacterial communities. However, the effectiveness of such combined technologies depends on the presence of soil microorganisms capable of degrading contaminants, making the use of genomic sequencing methods an important tool in the site characterization phase. Among these, 16 S rRNA amplicon sequencing stands out as a preferred, cost-effective tool for prokaryotic metataxonomic analysis in remediation practices, useful for identifying potential biodegrading bacteria at contaminated sites.

Conclusion

This study showed that heating to 100 ^oC severely affected the bacterial community, but it showed the ability to recover from the exposure to 60 ^oC. Network parameters remained stable, indicating that heating did not significantly alter the network topology, and that the complexity, efficiency, and resilience of the community network were preserved. However, heating to 60 ^oC resulted in a shift in dominance from Sphingobium to the BCP group after 55 days of incubation. Both genera are reported in the literature for their ability to degrade PAHs present in creosote, so we suggest that temperature does not affect biodegradation in soil, indicating that TEB can potentially be used for bioremediation. However, further studies are needed to determine the degradation rates of microorganisms under these conditions and to explore a wider range of temperatures in order to confirm that biodegradation processes are satisfactorily maintained.

Author contributions

Daniel Di Pace Penna Soares: soil sampling, formal analysis, investigation, methodology, writing original draft, review and editing. Valéria de Oliveira Maia: supervision, review, and editing. Juliana Gardenalli de Freitas: Funding acquisition, Project administration, review and editing. Kelly Johanna Hidalgo Martinez: bioinformatic analysis, review and editing. Alexandre Muselli Barbosa: resources, contaminated area management, review and editing. Cristina Rossi Nakayama: conceptualization, project administration, supervision, writing original draft, review and editing.

Funding

This work was supported by the National Council for Scientific and Technological Development, CNPq (Project number: 426953/2016-9); and The São Paulo Research Foundation, FAPESP (Project number: 2020/08164-6). The authors thank the technical team of the Institute for Technological Research for their support in the field work for the collection of soil samples.

Data availability

The datasets generated during and/or analysed during the current study are available in the EMBL Nucleotide Sequence Database (ENA) repository, https://www.ebi.ac.uk/ena/browser/view/PRJEB81236.

Declarations

Declaration of generative AI in scientific writing

The authors declare that they had no use of artificial intelligence (AI) or AI-assisted technologies in the scientific writing process of this paper.

competing of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.